Mitigating Stroke Risk in Type 2 Diabetes Through Physical Activity: A Nationwide Population-Based Study in Sweden
Article information
Abstract
Background and Purpose
Type 2 diabetes (T2DM) is associated with elevated stroke risk. Physical activity may attenuate this excess risk, but large-scale population evidence is limited. This study evaluated whether physical activity mitigates excess stroke risk in individuals with T2DM compared with age- and sex-matched individuals without diabetes.
Methods
All adults with T2DM registered in the Swedish National Diabetes Register between 2010 and 2019 were included and matched 1:3 to individuals without diabetes based on age and sex. Participants were stratified into five groups according to physical activity. Stroke incidence was estimated using Kaplan–Meier curves, and Cox proportional hazards models calculated adjusted hazard ratios (HRs) for ischemic and hemorrhagic stroke. Robustness was assessed using competing risks regression.
Results
The cohort included 369,704 individuals with T2DM (mean [standard deviation] age, 64.3 [12.3] years; 156,729 female individuals [42.4%]) and 1,109,112 matched individuals without diabetes. The adjusted HR for ischemic stroke in inactive individuals with T2DM compared with individuals without diabetes was 1.36 (95% confidence interval [CI] 1.23–1.50), declining with higher activity and becoming non-significant among those active ≥3 times/week (HR 1.05, 95% CI 0.96–1.15). The elevated ischemic stroke risk at higher glycated hemoglobin levels was attenuated in individuals with higher physical activity.
Conclusions
Higher physical activity levels were associated with reduced stroke risk in individuals with T2DM. Those active at least three times weekly had stroke risk comparable to individuals without diabetes of the same age and sex. Promoting physical activity should be an integral component of diabetes care and stroke prevention.
Introduction
Type 2 diabetes significantly increases the risk of ischemic stroke, overall cardiovascular disease, and premature death [1,2]. The association with hemorrhagic stroke remains less consistent, with studies reporting risks ranging from higher to lower compared with the general population [3], and some evidence suggesting a potential influence of glycemic control [2]. Physical activity reduces all-cause and cardiovascular morbidity and mortality [4], and interventions to increase physical activity can be broadly implemented with minimal resources [5]. Increasing total time spent in physical activity, even at light intensity, has been associated with reduced stroke risk in the general population [6]. Regular physical activity among individuals with diabetes is associated with improved cardiovascular risk factor control, including blood pressure, lipid profiles, body weight, and glycemic control [7]. Although limited, previous studies have suggested that higher physical activity levels are associated with reduced stroke risk and lower cardiovascular mortality among individuals with diabetes [8-13]. These studies have primarily examined differences within diabetes populations, without direct comparison with individuals without diabetes. Consequently, large-scale studies evaluating whether, and to what extent, the excess stroke risk associated with diabetes can be mitigated through physical activity remain limited.
In this study, we aim to investigate the association between physical activity and excess risk of ischemic and hemorrhagic stroke among individuals with type 2 diabetes, compared to ageand sex-matched controls from the general population.
Methods
Ethical approval
The study was approved by the Swedish Ethical Review Authority on August 18, 2021 (registration number: 2021–03645). Individual consent is not required for reporting patients to national health care quality registries under Swedish law (Patient Data Act 2008:355, chapter 7).
Data sources
All individuals with type 2 diabetes and a first registration in the Swedish National Diabetes Register (NDR) between January 1, 2010, and December 31, 2019, were included. The NDR collects detailed information on clinical and laboratory variables, diabetes care, and complications for most adults with diabetes in Sweden [14]. In 2019, register coverage was estimated at approximately 87% of adults with diabetes nationwide, indicating high completeness and representativeness [14]. For each diabetes case, three individuals without diabetes (defined as having no record in the NDR and no diabetes diagnosis [International Classification of Diseases, 10th Revision, ICD-10 E08–E13] in the National Patient Register during the study period) were randomly selected from the Swedish Total Population Register [15] and matched exactly on age and sex at the date of the index patient’s first NDR entry, which served as baseline.
Stroke outcomes and comorbidities were identified through the Swedish National Patient Register [16], which contains nationwide data on hospital admissions and outpatient specialist visits, with high sensitivity and positive predictive value for stroke diagnoses [17]. Information on deaths and causes of death was obtained from the Swedish Cause of Death Register [18]. Prescription medication data were retrieved from the Swedish Prescribed Drug Register [19]. Socioeconomic variables, including education and income, were obtained from the Longitudinal Integrated Database for Health Insurance and Labor Market Studies [20]. Data from these registers were collected for both individuals with diabetes and individuals without diabetes. All registers were linked by Statistics Sweden and the National Board of Health and Welfare using personal identity numbers assigned to all residents in Sweden at birth or shortly after immigration, and pseudonymized data were provided to the researchers.
Diabetes diagnoses
Type 2 diabetes diagnoses were defined using epidemiologic criteria. Individuals diagnosed at any age and managed with dietary modifications alone, oral glucose-lowering medications, or insulin (with or without oral agents) if diagnosed at age 40 or older, were classified as having type 2 diabetes. In the Swedish health care system, individuals with type 2 diabetes are generally managed in primary care.
Variables
Sociodemographic characteristics
Baseline sociodemographic characteristics included age, sex, education level during the previous year (categorized into 3 levels: lower ≤9 years, middle 10–12 years, higher >12 years), and family income during the previous year (in hundreds of Swedish Krona and categorized into 3 equal tertiles: lower <2,384, middle 2,384–4,420, higher >4,420). Baseline smoking status was obtained from the NDR and was available only for individuals with diabetes.
Comorbidities
Baseline comorbidities were identified using inpatient and outpatient diagnoses recorded for each individual from 4 years to 1 day before baseline, based on the ICD-10. Included conditions were cancer (C00–C96), dementia (F01–F03), alcohol or substance abuse (F10–F19), hypertension (I10–I15), atrial fibrillation (I48), myocardial infarction (I21–I23), heart failure (I50), TIA (G45.9), and prior stroke (ischemic [I63], hemorrhagic [I61], or unspecified type [I64]).
Medications
Baseline medications comprised all filled pharmacy prescriptions during the 6 months before and the 6 months after baseline registration. Included medication classes were antiplatelet agents, anticoagulants, antihypertensives, diuretics, lipid-lowering medications, and β-blockers/antiarrhythmics/glycosides.
Laboratory values
Baseline laboratory values were obtained from the first NDR registration and were available only for individuals with diabetes. Recorded measures included low-density lipoprotein, systolic and diastolic blood pressure, estimated glomerular filtration rate, body mass index, and glycated hemoglobin (HbA1c).
Physical activity
Physical activity was self-reported by individuals with diabetes and recorded repeatedly during routine clinical follow-up rather than as a single baseline measure. Follow-up visits are part of structured diabetes care but are not conducted at fixed intervals, resulting in variability in the timing and number of assessments across individuals. Physical activity was measured using a 5-level scale reflecting the frequency of engaging in at least 30 minutes of activity per week at an intensity comparable to outdoor walking. The categories were defined as follows: Level 1: Never; Level 2: <1 time per week; Level 3: 1–2 times per week; Level 4: 3–5 times per week; and Level 5: 6–7 times per week.
For each individual, all available physical activity registrations between January 1, 2010, and the date of censoring or stroke event were included. To summarize long-term physical activity levels while reducing the impact of occasional missing or atypical entries, the median physical activity level was calculated for each individual across this period, providing a more stable estimate of habitual activity over time. Individuals without any physical activity recordings before the censoring date were classified as missing.
Outcomes
The primary outcome was time to first occurrence of ischemic or hemorrhagic stroke, measured from the date of diabetes diagnosis recorded in the NDR. Stroke incidence was followed until death or censoring at 2 years after baseline. If both stroke types occurred, only the first event was considered. Follow-up time for each individual and matched controls began on the same date. Stroke events were identified using ICD-10 codes, including ischemic stroke (I63) and hemorrhagic stroke (I61). Strokes of unspecified type (I64) were classified as ischemic in accordance with Swedish stroke guidelines recommending immediate computed tomography imaging to exclude hemorrhage. Subarachnoid hemorrhage (I60) was excluded from the hemorrhagic stroke outcome.
Statistical analyses
Individuals with diabetes were grouped into five categories according to their median physical activity levels. Age- and sexmatched individuals without diabetes were assigned to the same physical activity category as their corresponding diabetes case. All subsequent analyses were stratified by physical activity level. Baseline descriptive characteristics were reported as means with standard deviations for continuous variables and as counts with valid percentages for categorical variables.
Stroke risk
Crude 2-year incidence of ischemic and hemorrhagic stroke was calculated as the proportion of individuals experiencing events within each group. Cumulative incidence was estimated using Kaplan–Meier curves with 95% confidence intervals (CIs).
Cox proportional hazards (PH) models stratified by physical activity level were used to estimate adjusted hazard ratios (HRs) for ischemic and hemorrhagic stroke in individuals with type 2 diabetes compared with individuals without diabetes. Separate models were constructed for ischemic and hemorrhagic stroke. In the ischemic stroke model, individuals were censored at death or hemorrhagic stroke occurring before an ischemic event. Similarly, in the hemorrhagic stroke model, individuals were censored at death or ischemic stroke. Time to event was measured from baseline to the first relevant stroke event, death, another stroke type, or end of follow-up 2 years after baseline. Covariates included age, sex, diabetes status, education, income, comorbidities, and medications.
The PH assumption was evaluated using scaled Schoenfeld residual plots, with separate assessments for each outcome and each level of stratification. All covariates met the assumption except prior stroke, which demonstrated a stronger effect early that diminished over time; however, it was retained because of its clinical importance and consistent direction of effect.
Competing risks analysis
Fine–Gray competing risks regression models were fitted to account for competing events. For ischemic stroke, hemorrhagic stroke, and death were treated as competing risks, whereas for hemorrhagic stroke, ischemic stroke, and death were considered competing events. The models were adjusted for the same covariates as the primary analysis.
HbA1c subgroup analysis
A subgroup analysis was performed using a second level of stratification based on baseline HbA1c levels, categorizing individuals with type 2 diabetes into five groups (≤42 mmol/mol, 43–52 mmol/mol, 53–62 mmol/mol, 63–72 mmol/mol, >72 mmol/mol). Matched individuals without diabetes were assigned to the same HbA1c category as their corresponding case. Cox PH models were fitted within each stratum for ischemic stroke only.
Missing data analysis
A complete-case analysis was performed, excluding individuals with missing physical activity data. To evaluate the characteristics and stroke risk of the excluded group, separate descriptive and Cox regression analyses were conducted for individuals with missing physical activity data compared with their corresponding individuals without diabetes. Among the remaining covariates, only education and income had very low proportions of missingness, and individuals with missing data on these variables were also excluded from the adjusted analysis (Supplementary Figure 1).
All statistical tests were two-tailed and performed at α=5%. All estimates are presented with 95% CIs. Statistical analyses were conducted using R (R Foundation for Statistical Computing, Vienna, Austria) and Statistical Package for the Social Sciences (version 30.0; IBM Corp., Armonk, NY, USA).
Results
Of 385,567 individuals with a first registration in the NDR between January 1, 2010, and December 31, 2019, 369,704 had type 2 diabetes according to the epidemiologic criteria (Supplementary Figure 1). The final study population consisted of 369,704 (25.0%) individuals with type 2 diabetes and 1,109,112 (75.0%) age- and sex-matched individuals without diabetes. Among individuals with type 2 diabetes who had recorded physical activity data (310,369), 36,790 (11.9%) reported never being physically active, 39,462 (12.7%) were active <1 time per week, 63,424 (20.4%) 1–2 times per week, 77,264 (24.9%) 3–5 times per week, and 93,429 (30.1%) 6–7 times per week. The median (interquartile range) number of recorded physical activity assessments per individual was 4 (2.0–6.0). During follow-up of up to 2 years per individual (2,874,915 person-years; mean [standard deviation, SD] follow-up 710.7 [95.7] days among individuals without diabetes and 706.3 [105.8] days among individuals with diabetes), 7,203 individuals with diabetes (2.0%) and 13,957 individuals without diabetes (1.3%) experienced a stroke. Descriptive characteristics of individuals with diabetes and individuals without diabetes are presented in Table 1. Descriptive characteristics stratified by physical activity level are presented in Supplementary Table 1. Individuals with diabetes who reported never engaging in physical activity were the oldest (mean age [SD], 66.3 [13.2] years) and had the highest prevalence of comorbid conditions (17,272 [46.9%] had at least one included comorbidity).
Descriptive characteristics (n=1,478,816)
Stroke risk
The 2-year crude incidence of ischemic stroke was highest among individuals with diabetes who reported never being physically active (2.5%) compared with matched individuals without diabetes (1.4%). The absolute difference between individuals with diabetes and individuals without diabetes narrowed as physical activity increased: 1.9% versus 1.1% (<1 time/week), 1.3% versus 1.0% (1–2 times/week), 1.0% versus 0.9% (3–5 times/week), and 1.2% versus 1.0% (6–7 times/week). The 2-year crude incidence of hemorrhagic stroke was higher among individuals with diabetes compared with individuals without diabetes among those who reported never being physically active (0.38% vs. 0.23%) and those active <1 time/week (0.23% vs. 0.17%). Among individuals with diabetes who reported higher levels of physical activity, the crude incidence of hemorrhagic stroke was similar to or lower than that of individuals without diabetes: 0.17% versus 0.18% (1–2 times/week), 0.14% versus 0.15% (3–5 times/week), and 0.14% versus 0.18% (6–7 times/week). Kaplan–Meier curves for ischemic and hemorrhagic stroke incidence by diabetes status, stratified by physical activity level, are presented in Figure 1. An interactive version of Figure 1, including 95% CIs, is available at: https://mavridis-et-al-interactive.pages.dev/.
Cumulative incidence of ischemic (A) and hemorrhagic stroke (B) over 2 years, stratified by physical activity level and diabetes status. Interactive plots are available at: https://mavridis-et-al-interactive.pages.dev/.
Figure 2 summarizes adjusted HRs for the relative 2-year risk of ischemic and hemorrhagic stroke across physical activity levels among individuals with type 2 diabetes compared with individuals without diabetes. Full Cox model outputs by activity level are provided in Supplementary Table 2 (ischemic stroke) and Supplementary Table 3 (hemorrhagic stroke). Relative to individuals without diabetes, individuals with type 2 diabetes who reported no physical activity had the greatest risk of ischemic stroke (HR: 1.36; 95% CI 1.23–1.50), followed by those active <1 time/week (HR: 1.31; 95% CI 1.18–1.46). The excess risk decreased with increasing activity and was not statistically significant among those reporting physical activity ≥3 times per week. For hemorrhagic stroke, individuals with diabetes who reported no physical activity had a higher risk than individuals without diabetes (HR: 1.33; 95% CI 1.03–1.72). No statistically significant association was observed in the “<1 time/week,” “1–2 times/week,” or “3–5 times/week” groups. In the most active group (“6–7 times/week”), individuals with diabetes had a lower risk of hemorrhagic stroke than individuals without diabetes (HR: 0.74; 95% CI 0.60– 0.92). Competing risks analyses were consistent with the primary results showing similar associations between diabetes status, physical activity, and stroke outcomes after accounting for competing events (Supplementary Table 4).
Forest plot of HRs and 95% CIs for ischemic (A) and hemorrhagic stroke (B) associated with diabetes, stratified by physical activity level. Models were adjusted for age, sex, comorbidities, previous stroke, medications, education, and income. Reference category: individuals without diabetes. HR, hazard ratio; CI, confidence interval.
HbA1c subgroup analysis
In the subgroup analysis stratified by physical activity and HbA1c levels, the risk of ischemic stroke increased progressively with higher HbA1c, particularly among individuals with lower physical activity. Among individuals with diabetes who reported no physical activity, HRs ranged from 1.20 (95% CI 0.90–1.59) in the HbA1c ≤42 mmol/mol group to 2.19 (95% CI 1.64–2.93) in the HbA1c >72 mmol/mol group. In contrast, individuals reporting physical activity ≥3 times per week demonstrated weaker and less consistent associations, with statistically elevated risk confined to the higher HbA1c categories. No significant excess risk of ischemic stroke was observed in the lowest HbA1c category (≤42 mmol/mol) at any physical activity level (Figure 3).
Forest plot of HRs and 95% CIs for stroke associated with diabetes, stratified by physical activity level and HbA1c category. Models were adjusted for age, sex, comorbidities, previous stroke, medications, education, and income. Reference category: individuals without diabetes. HbA1c, glycated hemoglobin; HR, hazard ratio; CI, confidence interval.
Missing data analysis
Because of missing physical activity data, 59,335 individuals were excluded from the primary analysis (Supplementary Figure 1). These individuals were older (mean age [SD], 67.0 [13.9] vs. 63.7 [12.0] years) and had a greater burden of comorbidities (26,424 [44.5%] vs. 112,836 [36.4%] with at least one comorbid condition) compared with those with complete data (Supplementary Table 5). In Cox regression analyses (Supplementary Table 6), individuals with diabetes and missing physical activity data had a 92% higher risk of ischemic stroke compared with individuals without diabetes (HR: 1.92; 95% CI 1.79–2.05) and a 31% higher risk of hemorrhagic stroke (HR: 1.31; 95% CI 1.09–1.58).
Discussion
This nationwide observational cohort study investigated whether physical activity mitigates the excess stroke risk associated with type 2 diabetes. Using age- and sex-matched individuals without diabetes from the general population and stratifying analyses by physical activity level, we observed that individuals with type 2 diabetes who were physically inactive had a higher risk of both ischemic and hemorrhagic stroke compared with individuals without diabetes. This excess risk decreased progressively with increasing physical activity. Individuals with diabetes who engaged in physical activity for at least 30 minutes on three or more days per week had no elevated risk of ischemic stroke compared with age- and sex-matched individuals without diabetes. For hemorrhagic stroke, increased risk was confined to individuals with diabetes who were physically inactive. Notably, in the most active group, individuals with diabetes had a lower risk of hemorrhagic stroke than their matched counterparts without diabetes. Although the benefits of physical activity are well established, prior studies have primarily examined differences within diabetes populations without directly comparing stroke risk to individuals without diabetes [8-13]. The present study addresses this gap by demonstrating that regular physical activity can reduce diabetes-associated excess stroke risk to a level comparable with that observed in the general population.
Across physical activity levels, consistent patterns in baseline characteristics emerged. Overall, males with diabetes were more likely to report higher levels of physical activity. This pattern has been previously reported in the general population [21], as well as among individuals with diabetes [13,22]. Physical activity declined with increasing age; however, the most active group demonstrated a modest increase in mean age compared with the preceding activity groups, a pattern previously described [22]. The prevalence of most comorbidities decreased with increasing physical activity, suggesting a more favorable health profile among more active individuals. Similar associations have been reported previously [22]. Both education and income increased progressively across physical activity levels, consistent with prior findings [13,22].
For ischemic stroke, a graded reduction in excess risk across physical activity levels was observed. Individuals with type 2 diabetes in the “Never” and “<1 time per week” groups had a 36% and 31% higher risk of ischemic stroke, respectively, compared with their matched counterparts without diabetes. In the “1–2 times per week” group, the excess risk decreased to 10%, and no statistically significant excess risk remained in the “3–5” and “6–7 times per week” groups. Previous studies have shown that higher levels of physical activity are associated with lower stroke risk among individuals with type 2 diabetes [8-13]. However, to our knowledge, no prior study has evaluated whether physical activity attenuates the excess stroke risk attributable to diabetes compared with individuals without diabetes. The observed risk reduction likely reflects both the physiological effects of regular activity and correlated health behaviors, including improved cardiovascular risk factor control and adherence to preventive care.
For hemorrhagic stroke, prior research in individuals with type 2 diabetes has produced inconsistent findings [3], with some studies reporting increased risk [23], others suggesting decreased risk [24], and some showing no significant association [2]. Although physical activity is generally associated with lower hemorrhagic stroke risk in the general population [25,26], evidence specifically evaluating its effect in individuals with type 2 diabetes remains limited and has not demonstrated a clear protective association [26]. In the present study, individuals with diabetes in the “Never” group had a 33% higher risk of hemorrhagic stroke compared with their matched counterparts without diabetes, whereas no statistically significant excess risk was observed in the “<1 time/week,” “1–2 times/week,” and “3–5 times/week” groups. In the “6–7 times/week” group, individuals with diabetes had a lower risk of hemorrhagic stroke than individuals without diabetes. These findings suggest a potential protective association of high-frequency physical activity with hemorrhagic stroke risk in individuals with diabetes; however, further investigation is warranted to confirm this relationship and clarify underlying mechanisms.
The present findings indicate that individuals with diabetes who engage in 30-minute sessions of physical activity at least 3–5 times per week have a stroke risk comparable with that of age- and sex-matched individuals without diabetes. This observation aligns with current European and American diabetes management guidelines recommending a minimum of 150 minutes of moderate-intensity physical activity per week [27,28], and reinforces the role of regular physical activity in reducing stroke risk among individuals with diabetes.
Previous studies have consistently demonstrated a strong association between elevated HbA1c levels and increased risk of ischemic stroke [29]. Consistent with this evidence, we confirmed that higher HbA1c levels are associated with greater stroke risk. Importantly, our findings further indicate that higher levels of physical activity attenuate this association. Individuals in the lowest HbA1c category (≤42 mmol/mol) did not exhibit an increased risk of stroke at any physical activity level, whereas those in the highest category (>72 mmol/mol) had a higher risk across all activity levels. These findings emphasize the central role of glycemic control in stroke prevention and suggest that physical activity may mitigate, although not fully eliminate, excess stroke risk associated with poor glycemic control.
In the missing data analysis, individuals with type 2 diabetes who lacked physical activity information were older and had a greater burden of comorbidities than those with complete data. Relative to their matched counterparts without diabetes, these individuals had a 92% higher risk of ischemic stroke and a 31% higher risk of hemorrhagic stroke. These estimates were comparable to or exceeded those observed in the “Never” physical activity group. This pattern suggests that missing physical activity data may not be random and may reflect poorer underlying health status. It is also plausible that physical activity is less frequently assessed or documented in individuals with more complex medical needs because of competing clinical priorities or assumptions regarding limited activity capacity. These observations align with prior research indicating that non-response to lifestyle questions, including physical activity, is more common among higher-risk subgroups [30].
Strengths and limitations
Strengths of this study include its longitudinal design and the use of large, population-representative, register-based data with near-complete coverage of individuals with type 2 diabetes in Sweden. The large sample size enabled subgroup analyses and estimation of stroke risk across physical activity levels. The matched cohort design, combined with linkage to high-quality national registers, allowed comprehensive adjustment for a broad range of covariates, thereby enhancing internal validity and reducing confounding. In addition, the use of both cause-specific and competing risks models strengthened robustness across different modeling assumptions and accounted for the influence of competing events. However, several clinical and laboratory covariates were obtained from the NDR and were therefore unavailable for individuals without diabetes. The use of repeated physical activity measurements provided a more stable estimate of habitual activity compared with a single baseline assessment. Nevertheless, physical activity was self-reported, introducing the possibility of recall and social desirability bias. Moreover, summarizing exposure using median values may obscure meaningful changes in physical activity over time. Physical activity data were missing for a proportion of individuals, and those with missing data had a higher-risk profile, suggesting that physical activity may be less frequently assessed or documented in individuals with more complex medical conditions. Additionally, because of limited event counts, individuals with type 1 diabetes were excluded from the analysis. It cannot be excluded that the epidemiologic definition of type 2 diabetes resulted in some misclassification of diabetes type. Subgroup analysis by HbA1c level was not performed for hemorrhagic stroke because the number of events within strata was insufficient for reliable estimation. Finally, as in all observational studies, residual confounding and unmeasured variables may persist, and causal inference cannot be established. Therefore, these findings should be interpreted cautiously but may be generalizable to other Western countries with publicly funded health care systems comparable to Sweden, while recognizing potential differences in health care delivery, population characteristics, and registry infrastructure.
Conclusions
In this nationwide cohort study, higher levels of physical activity were associated with a graded reduction in excess risk of both ischemic and hemorrhagic stroke among individuals with type 2 diabetes. These findings indicate that regular physical activity may attenuate stroke risk in this high-risk population. Promoting physical activity should remain a central component of diabetes management for both metabolic regulation and cardiovascular risk reduction. Future research should evaluate how longitudinal changes in physical activity patterns influence stroke risk in individuals with type 2 diabetes and examine strategies to facilitate sustained engagement in regular physical activity within routine clinical care.
Supplementary materials
Supplementary materials related to this article can be found online at https://doi.org/10.5853/jos.2025.05113.
Descriptive characteristics for individuals with diabetes and controls by level of physical activity
Cox regression results for ischemic stroke stratified by physical activity level
Cox regression results for hemorrhagic stroke stratified by physical activity level
Competing risks regression results for ischemic and hemorrhagic stroke association with type 2 diabetes stratified by physical activity level
Descriptive characteristics for individuals with diabetes and controls by missing physical activity level
Cox regression results for ischemic and hemorrhagic stroke for individuals with missing physical activity level
Flowchart of the study sample. NDR, National Diabetes Register.
Notes
Funding statement
This study was supported by grants from Sahlgrenska University Hospital funds (SU-997998 and SU-1016425), the Swedish National Stroke Association, the Swedish Brain Foundation, the Swedish Heart Lung Foundation, and the Swedish state under the ALF agreement between the Swedish government and the county councils (ALFGBG-965653).
Conflicts of interest
The authors have no financial conflicts of interest.
Author contribution
Conceptualization: Anastasios Mavridis, Dongni Buvarp, Adam Viktorisson. Study design: Anastasios Mavridis, Dongni Buvarp, Adam Viktorisson. Methodology: Anastasios Mavridis, Dongni Buvarp, Adam Viktorisson. Data collection: Adam Viktorisson. Investigation: Anastasios Mavridis. Statistical analysis: Anastasios Mavridis. Writing—original draft: Anastasios Mavridis. Writing— review & editing: all authors. Funding acquisition: Adam Viktorisson. Approval of final manuscript: all authors.
Acknowledgments
The authors express their gratitude to the study participants and the clinical support staff for their contributions to data collection during routine clinical care. The authors also thank Professor Katharina Stibrant Sunnerhagen for funding support.
According to Swedish regulations (https://etikprovningsmyndigheten.se/en/what-the-act-says/), data may only be used as approved in the ethical application for this study. Qualified researchers may request access to the dataset by contacting the authors (contact Dr. Adam Viktorisson, email: adam.viktorisson@gu.se).